Creating open resources for an introductory data science course

Abstract

Open-source educational resources have many benefits for both instructors and learners alike, such as cost-savings for learners, quick iteration on materials for instructors, and increased quality through auditable collaboration. And given that the field of data science has embraced open-source development practices in such a big way, we believe that data science educators should "practice what they preach" by teaching with, and sharing, open-source tools and resources. However, data science is a very new field; educational resources at the truly introductory level are sparse and generally not based on real classroom experience. In this talk, we will discuss an open-source textbook (https://datasciencebook.ca) and complementary worksheets for introducing students to data science that we have created at the University of British Columbia. These resources were built using rigorous open-source workflows and tools (e.g., R, bookdown, Python, Jupyter, Git, and GitHub) in a public repository (https://github.com/ubc-dsci/introduction-to-datascience), and have been tested in a first-year undergraduate course that has served thousands of students over three years. Specifically, we will comment on the initial objectives of the project and how they evolved, tooling choices we made and their rationale, our collaboration strategy, and hard-won lessons we learned both in-class and offline about effective data science pedagogy in a large-scale classroom.

• Slides